Customized nutrient management of onion (Alium cepa) agroecosystems

Abstract

While onion cultivars, irrigation and soil and crop management practices have been given much attention in Brazil, nutrient management at growers’ scale is still challenging. Our objective was to customize the fertilization of onion crops. We attempted to adjust nutrient management to the complexity of onion cropping systems by combining ML and compositional methods. We assembled climatic, edaphic, and managerial features as well as tissue tests into a data set of 1182 observations collected across fertilizer experiments conducted over 13 years. Data were processed using machine learning methods. Fertilization (NPK) treatments as well as edaphic and managerial features that are easy to acquire by stakeholders sufficed to explain 93.5% of total variation in marketable onion yields. Customized crop response models differed from state-base fertilizer recommendations, indicating potential benefits to customize fertilizer recommendations using a median experimental site condition in southern Brazil. Foliar nutrient standards to reach > 50 Mg bulb ha− 1 differed among cultivars grown under a large range of edaphic and managerial features, supporting local nutrient diagnosis. Larger and more diversified observational and experimental data sets could be acquired to customize fertilization across more Brazilian onion agroecosystems and document successful combinations of growth-impacting features through close ethical collaboration among stakeholders.

Optimizing the Nitrogen Use Efficiency in Vegetable Crops

“…Data required to develop adequate nutrient budgets and calibration curves include total N uptake amount over the growing season (Table 6); patterns of N uptake, based on crop phenological stage; tissue N levels during particular periods of crop growth; and N residual levels or mineralization rates for representative soils in the region. A challenge to develop appropriate N fertilizer recommendations for vegetable crops is that calibration data or N response curves are location- [136,138], crop species-, cultivar-, and production system-specific (such as plasticulture vs. bare-ground culture, or conventional vs. organic). Calibration research should also be revisited periodically, as new cultivars are continually introduced, as production practices are adjusted, and as new technology is adopted to improve production efficiency.…”